scrypt was designed to use more memory. We can use its parameters to specify (approximately) how much memory we want it to require.
But both scrypt and PBKDF2 allow us to specify the output data size.
So my question is, can't we also make PBKDF2 use a lot of memory by simply asking it for a large output (and taking the bytes we actually need from the end of output, discarding the rest)?
If this is the case, then what's the point of scrypt?
If this is not the case (which is more likely), please tell me what's wrong with my thinking.